Category: Technology

On Tuesday, at its annual Snapdragon Summit in Hawaii, Qualcomm announced a major addition to its line of mobile chips with the Snapdragon X Elite, which the company is calling its most powerful processor to date. 

The Arm-based Snapdragon X Elite is the successor to last year’s Snapdragon 8cx Gen 3 line of laptop chips, which recently got a name change to reflect the huge leap in performance for this upcoming generation. Powered by 12 Oryon cores, Qualcomm claims the X Elite provides up to two times faster CPU performance compared to Intel’s 13th-gen Core i7-1360P and i7-1355U processors while also drawing up to 68 percent less power. 

The chip is based on a 4nm design fabricated by TSMC with standard clock speeds of 3.8GHz with a dual-core boost of up to 4.3GHz. Qualcomm also includes 42MB of total cache with an LPDDR5x memory bandwidth of 136 GB/s. When compared to Apple’s M2 chip, Qualcomm says the X Elite boasts 50 percent faster peak multithreaded performance. And thanks to its integrated GPU, the X Elite is said to offer twice the graphics performance as the i7-13800H at ISO power. 

Qualcomm’s new AI Engine should also provide a significant boost to machine learning-based tasks. Between its Oryon CPU, Adreno GPU and Hexagon NPU, the X Elite looks to deliver up to 75 TOPs, which Qualcomm claims is 4.5 times more than its competitors. The chip was also designed to run large language models with up to 13 billion parameters locally with Qualcomm saying the chip offers the fastest Stable Diffusion performance of any laptop chip on the market. Other features include support for AV1 4K HDR video encoding/decoding, 5G connectivity (with downloads of up to 10 Gbps), Wi-Fi 7 and the company’s built-in Sensing Hub.

The arrival of the Snapdragon X Elite is poised to be a major breakthrough for Qualcomm as it attempts to better compete against rivals like Intel and more specifically Apple, following the latter’s transition away from x86-based chips to its custom-designed M-series silicon. The development of the chip’s Oryon cores has been a multiyear process following Qualcomm’s acquisition of Nuvia back in 2021. 

That said, while the X Elite boasts some impressive specs and power numbers, the major challenge for Qualcomm is translating all that oomph into actual performance. That’s because unlike Apple, Qualcomm doesn’t have the luxury of being able to design both its chips and the software/OS its processors are running. And as we’ve seen in the past on devices like the Surface Pro 9, systems running Windows on Arm still don’t feel as fast or responsive when compared to their more traditional x86-based counterparts. 

Thankfully, with retail PCs powered by the Snapdragon X Elite expected to arrive sometime in mid-2024, it won’t be too long until we can see how Qualcomm’s latest laptop chip performs in the real world. 

This article originally appeared on Engadget at https://www.engadget.com/the-snapdragon-x-elite-is-qualcomms-most-powerful-chip-to-date-190004830.html?src=rss 

Qualcomm is no stranger in running artificial intelligence and machine learning systems on-device and without an internet connection. They’ve been doing it with their camera chipsets for years. But on Tuesday at Snapdragon Summit 2023, the company announced that on-device AI is finally coming to mobile devices and Windows 11 PCs as part of the new Snapdragon 8 Gen 3 and X Elite chips.

Both chipsets were built from the ground up with generative AI capabilities in mind and are able to support a variety of large language models (LLM), language vision models (LVM), and transformer network-based automatic speech recognition (ASR) models, up to 10 billion parameters for the SD8 gen 3 and 13 billion parameters for the X Elite, entirely on-device. That means you’ll be able to run anything from Baidu’s ERNIE 3.5 to OpenAI’s Whisper, Meta’s Llama 2 or Google’s Gecko on your phone or laptop, without an internet connection. Qualcomm’s chips are optimized for voice, text and image inputs.

“It’s important to have a wide array of support underneath the hood for these models to be running and therefore heterogeneous compute is extremely important,” Durga Malladi, SVP & General Manager, Technology Planning & Edge Solutions at Qualcomm, told reporters at a prebriefing last week. “We have state-of-the-art CPU, GPU, and NPU (Neural Processing Unit) processors that are used concurrently, as multiple models are running at any given point in time.”

The Qualcomm AI Engine is comprised of the Oryon CPU, the Adreno GPU and Hexagon NPU. Combined, they handle up to 45 TOPS (trillions of operations per second) and can crunch 30 tokens per second on laptops, 20 tokens per second on mobile devices — tokens being the basic text/data unit that LLMs can process/generate off of. The chipsets use Samsung’s 4.8GHz LP-DDR5x DRAM for their memory allocation.

“Generative AI has demonstrated the ability to take very complex tasks, solve them and resolve them in a very efficient manner,” he continued. Potential use cases could include meeting and document summarization or email drafting for consumers, and prompt-based computer code or music generation for enterprise applications, Malladi noted.

Or you could just use it to take pretty pictures. Qualcomm is integrating its previous work with edge AI, Cognitive ISP. Devices using these chipsets will be able to edit photos in real-time and in as many as 12 layers. They’ll also be able to capture clearer images in low light, remove unwanted objects from photos (a la Google’s Magic Eraser) or expand image backgrounds. User scan even watermark their shots as being real and not AI generated, using Truepic photo capture.

Having an AI that lives primarily on your phone or mobile device, rather than in the cloud, will offer users myriad benefits over the current system. Much like enterprise AIs that take a general model (e.g. GPT-4) and tune it using a company’s internal data to provide more accurate and on-topic answers, a locally-stored AI “over time… gradually get personalized,” Malladi said, “in the sense that… the assistant gets smarter and better, running on the device in itself.”

What’s more, the inherent delay present when the model has to query the cloud for processing or information doesn’t exist when all of the assets are local. As such, both the X Elite and SD8 gen 3 are capable of not only running Stable Diffusion on-device but generating images in less than 0.6 seconds.

The capacity to run bigger and more capable models, and interact with those models using our speaking words instead of our typing words, could ultimately prove the biggest boon to consumers. “There’s a very unique way in which we start interfacing the devices and voice becomes a far more natural interface towards these devices — as well in addition to everything else,” Malladi said. “We believe that it has the potential to be a transformative moment, where we start interacting with devices in a very different way compared to what we’ve done before.”

Mobile devices and PCs are just the start for Qualcomm’s on-device AI plans. The 10-13 billion parameter limit is already moving towards 20 billion-plus parameters as the company develops new chip iterations. “These are very sophisticated models,” Malladi commented. “The use cases that you build on this are quite impressive.”

“When you start thinking about ADAS (Advanced Driver Assistance Systems) and you have multi-modality [data] coming in from multiple cameras, IR sensors, radar, lidar — in addition to voice, which is the human that is inside the vehicle in itself,” he continued. “The size of that model is pretty large, we’re talking about 30 to 60 billion parameters already.” Eventually, these on-device models could approach 100 billion parameters or more, according to Qualcomm’s estimates.

This article originally appeared on Engadget at https://www.engadget.com/qualcomm-brings-on-device-ai-to-mobile-and-pc-190030938.html?src=rss 

At its annual Snapdragon Summit on Tuesday, Qualcomm revealed its latest mobile chipset. Perhaps the biggest change in the Snapdragon 8 Gen 3 is the introduction of on-device generative AI (akin to Google’s Tensor G3). The chipset’s AI Engine supports multi-modal generative AI models and what Qualcomm claims is the world’s fastest Stable Diffusion system with the ability to generate an image in under a second. So, you should be able to whip up backgrounds and images for social media posts in a flash.

Because GAI requests are handled on-device, Qualcomm says they remain private. There’s an AI assistant based on Meta’s Llama 2 language learning model. It can securely use personal data such as your favorite activities, location and fitness level to provide more personalized responses, Qualcomm says. The Qualcomm Sensing Hub (which powers this feature) is said to deliver an AI performance increase of up to 3.5 times that of Qualcomm’s previous mobile chipset. The company also claims the Snapdragon 8 Gen 3 delivers up to 98 percent faster Hexagon NPU performance and 40 percent better performance per watt.

The camera and editing systems include support for features such as the ability to remove people and objects from a video with the tap of a finger. Qualcomm is promising voice-activated GAI photo and video editing thanks to its Cognitive ISP, along with the option to optimize different parts of a photo in up to 12 layers, whether you capture it with the front or back camera.

The Photo Expansion feature uses GAI to extend the borders of a photo, akin to Adobe’s Generative Fill tool. The Vlogger’s View mode allows you to share video from both your rear and front-facing cameras at the same time, so you can feature your face in what you capture as you narrate what you’re seeing.

Qualcomm says that Dolby HDR tech allows for image playback and capture with a fuller range of colors, tones and shades. Meanwhile, you can apply a cryptographic seal from Truepic to a photo to indicate that it’s the real deal and not something you whipped up with GAI, because that’s the world we’re living in these days. It’s a different approach to applying a label to AI-generated content.

Elsewhere, Qualcomm is promising some gaming upgrades in the Snapdragon 8 Gen 3. There’s support for Unreal Engine 5.2 along with hardware-accelerated ray-tracing with global illumination, which the company says is a first for a mobile chipset. This, Qualcomm says, helps deliver “lifelike, multi-source lighting” in games.

The Adreno Frame Motion Engine 2.0 can double framerates while sustaining the same power consumption as the previous chipset, Qualcomm claims. If you’re so inclined to hook up a Snapdragon 8 Gen 3-powered device to an external 240Hz display, you’ll be able to play games at up to 240 fps. 

Snapdragon Game Super Resolution, Qualcomm’s answer to upscaling tech like NVIDIA’s DLSS, will help games to run at up to 8K on supported displays too. The company notes that it’s open-sourcing Snapdragon Game Super Resolution.

You can expect gameplay audio to be synced to the millisecond, according to Qualcomm, meaning that you shouldn’t experience any lag while using earbuds. It’s also promising uninterrupted lossless audio even as you move away from your phone or from one room to another thanks to its Expanded Personal Area Network Technology system. The Snapdragon 8 Gen 3 is said to support 24-bit 96 kHz lossless music over Bluetooth.

On the subject of connectivity, Qualcomm is using AI hardware acceleration in the X75 Modem-RF System. It says this can help to deliver better 5G speeds, coverage, mobility, link robustness and location accuracy. There’s Wi-Fi 7 support too.

Compared with the previous chipset, Qualcomm says the Snapdragon 8 Gen 3’s Kryo CPU delivers 30 percent better performance and 20 percent more power efficiency. As for the GPU, it claims you’ll benefit from a 25 percent performance improvement and 25 percent greater power efficiency, along with a 40 percent ray-tracing upgrade. Overall, the Snapdragon 8 Gen 3 is slated to have 10 percent overall power savings compared with last year’s chip.

You won’t have to wait long to try out a phone with the Snapdragon 8 Gen 3 if you really want to. Qualcomm says devices that use the chipset should start to appear in the coming weeks. Among the brands and manufacturers that will use it are Sony, ASUS, OnePlus, Oppo, Vivo, Xiaomi, Honor and ZTE.

This article originally appeared on Engadget at https://www.engadget.com/qualcomms-snapdragon-8-gen-3-brings-on-device-generative-ai-to-more-android-phones-190019288.html?src=rss 

Apple just confirmed another streaming event for October 30 at, wait for it, 8PM ET. Tis the spooky season, so why not hold a presentation at night? The tagline for the event is “Scary Fast” so it certainly sounds like there will be some hardware announcements. We already had the big iPhone 15 event, so that leaves the company’s desktop and laptop computers.

To that end, it’s been a surprisingly long time since there’s been a new iMac. Analyst Mark Gurman suggested that this event will feature a refresh of the aging iMac line, and his track record is pretty good when it comes to Apple. Gurman also suggested that it’s likely we’ll see an update to the MacBook Pro.

This analysis is based on current retail supplies and shipping dates for forthcoming models, with retail stores noting they have short supply of the current iMacs and the 13-inch MacBook Pro. The colorful 24-inch M1 iMac came out all the way back in April 2021, which is a lifetime in tech circles, and hasn’t been updated since. The MacBook Pro M2 is newer, releasing in 2022, but still due for an upgrade.

With that said, there are also rumors that the event will reveal the followup chip, the M3. It’s also possible we’ll get an iPad reveal, as the company’s flagship tablet is also due for a refresh.

In the past, Apple has typically reserved its October press event for computers, which is another mark in the “new iMac” column. In any event, we don’t have that long to wait, as October 30 is just around the corner. Here’s hoping Tim Cook will be dressed as a mummy or Barbenheimer or something.

This article originally appeared on Engadget at https://www.engadget.com/apple-confirms-scary-fast-event-for-the-night-of-october-30-164836214.html?src=rss